Back to search
Top Contributors
1
commits
Vector compression with TurboQuant codecs for embeddings, retrieval, and KV-cache. 10x compression, pure NumPy core — optional GPU acceleration via PyTorch (CUDA/MPS) or MLX (Metal).
Stars
17
Forks
1
Watchers
17
Open Issues
0
Overall repository health assessment
No package.json found
This might not be a Node.js project
1
commits
add compute backend abstraction layer for GPU acceleration
4c9d10fView on GitHub